Some of my colleagues at work today drew my attention to what is now my all-time favourite Nature paper (and yes, I mean that in the most cynical possible way). In the article "Athletics: Momentous sprint at the 2156 Olympics?" [Nature 431, 525 (2004)], the authors plot both male and female 100m Olympic sprint times against year, claim that a linear interpolation represents the best possible fitting, and conclude by extrapolation that in the year 2156 females will overtake males. And no, that's not a typo, this was published in the highly reputable journal Nature.
Today I am proud to present my follow up results to this significant finding, based on the same data and techniques. Based on an extrapolation of the linear regression, I calculate that by the year 3027 the men’s Olympic 100m record will be 0.0 seconds. Of course, by this time there will be no 100m women’s event, since they will all be finishing the race before it starts. It follows from this extrapolation that at some stage in the late 2900's, we can expect Einstein's special theory of relativity to break down, paving the way for exciting new research into fundamental physics. I wish I could be around for that. Then again, if Einstein’s special theory of relativity breaks down, maybe I can be.
In conclusion, don't judge articles by the journal in which they're published. Shame of you Nature for publishing this dribble.